The proposed study builds upon current research addressing the inter- relationships among chronic impairment, disability and depression in later life and will make several unique contributions. The study's focus is on subgroup of elders experiencing an extremely common age- related disability, vision impairment, in which the rates of both functional disability and depression are particularly high, but which has received relatively little systematic research attention. The longitudinal design will permit an in-depth examination of the course of depression over time, utilizing a stress and coping conceptual model which incorporates key personal and social resources as mediators of the relationship between disability and depression. Most importantly, the focus on a sample seeking rehabilitation permits us to empirically challenge the assumption of inevitable, reciprocal decline in functional and depressive status. By following the natural course of both depression and rehabilitation service use, a primary long-term goal is to examine the extent to which, and mechanisms by which, this non- psychiatric intervention may influence depression status among disabled elders. This knowledge will provide a foundation for future intervention studies. The specific study aims are: 1. to document the prevalence, course, and severity of depression among visually impaired elders over time. 2. to examine the influence of depression on utilization of vision rehabilitation services. 3. to examine the mechanisms by which vision rehabilitation services may affect the severity and course of depression. 4. to test a longitudinal model explicating the interrelationships among vision impairment severity, co- morbid health conditions, functional disability, rehabilitation service utilization and depression, and how such relationships are mediated by personal and social resources. 5. to examine gender differences relative to aims 1-4 above. 600 elders (300 of each gender) will be sampled from applicants, age 65+, of a vision rehabilitation agency. Subjects will be assessed 4 times (at baseline, 6, 12 and 18 months) in order to examine both short and long-term causal relationships among key variables. Analyses address concurrent associations (cross-sectional) and prospective relationships (longitudinal) using regression and structural equation modeling (SEM) techniques.